Skip to main content

Welcome to Zingle AI

Zingle AI is an AI copilot for dbt data teams. It connects to your dbt GitHub repositories and data warehouse and lets analysts do end-to-end analytics engineering through conversational AI - with a human reviewing every change before it merges.

Where dbt gives you the framework to build reliable, version-controlled data transformations, Zingle AI gives you an agent that understands your model graph and semantic layer, writes and modifies SQL + YAML, validates with dbt compile/dbt test, estimates cost, and opens pull requests for you.

New here?

Follow the Onboarding guide to connect your first repository and warehouse, then start building in Studio.

What can it do?​

🛠️ Studio - Build agent

Conversationally create and modify dbt models. The agent writes SQL and schema YAML on a dedicated branch, validates it, shows diffs, and opens a PR.

Explore Studio →

♻️ Refactor pipelines

Multi-phase, DAG-driven automated repo refactors - merge/dedup/rename models, re-layer projects, build semantic models - with per-task review.

Run a refactor →

⚡ Optimize scans

Scan an existing repo for performance, materialization, DRY, and testing issues, generate fix candidates, apply them, and report cost impact.

Optimize your project →

🔎 Data Explorer

Browse the synced dbt model catalog, semantic layer, lineage graphs, and live warehouse tables - all in one place.

Explore your catalog →

🤖 Agent Builder

Create and configure custom agents - their prompt, tools, skills, and model - plus reusable skills. The platform's extensibility surface.

Build an agent →

How it fits together​

Zingle AI connects to two systems you already own - your dbt Git repository and your data warehouse - and adds an AI layer on top:

  • It reads your dbt project and warehouse metadata to build a context-aware catalog (models, columns, lineage, and the semantic layer).
  • Its agents write and modify SQL and YAML, validate changes with dbt, and estimate warehouse cost.
  • Every change lands as a pull request you review and merge - Zingle never merges to your default branch on its own.

The typical journey​

  1. Sign in with a passwordless email code.
  2. Connect GitHub by installing the Zingle AI GitHub App.
  3. Add a repository and a warehouse connection.
  4. Create an environment binding the repo to a warehouse target.
  5. A background sync parses your dbt project and populates the catalog.
  6. Use Studio, Refactor, Optimize, and the Data Explorer.

Ready? Head to the Onboarding guide.